| Topics: |
|
Při povídání o geometrii budeme kombinovat geometrický a algebraický přístup. Přičemž algebraický bude převažovat, jelikož umožňuje efektivní odvození výpočetních metod. Je tedy vhodné osvětžit si termíny jako lineární prostor, báze, homogenní soustava rovnic, skalární a vektorový součin, lineární zobrazení, lineární nezávislost vektorů, ...
It may help to get quickly acquainted with Matlab for programming for images. All functions are accompanied by ready-to-run demos which facilitate the start.
Šonka, Milan and Hlaváč, Václav and Boyle, Roger. Image Processing, Analysis and Machine Vision. 2007. Thomson. 3rd edition.
A thick book about general image processing and computer vision. We will use only few sections.
Hartley, Richard and Zisserman, Andrew. Multiple view geometry in computer vision. 2003. Cambridge University, Cambridge. 2nd edition.
The Book about 3D geometry in computer vision. Clean explanation, accessible for newcomers. We will use mainly chapters 4,6,7,9,11,12. The introductory chapters 2,3 explain the projective geometry we will need.
The paper presents an algebraic approach for calibration from unknown rotations. It is based on computation of pairwise homographies. It Although written long ago the paper uses modern notation which was adopted by many and it is thus very readable.
Wang, Lei and Kang, Sing Bing and Shum, Heung-Yeung and Xu, Guangyou. Error Analysis of Pure Rotation-Based Self-Calibration. 2004. IEEE Transactions on Pattern Analysis and Machine Intelligence, 26(2):275-280.
Clear description of various calibration cases. Errors due to translation.
Hansen, Michael and Sommer, Gerald. Active Depth Estimation with Gaze and Vergence Control using Gabor Filters. 1996. in 13th International Conference on Pattern Recognition.
Section 5 is the most relevant for us. The computer vision part may be skiped for the first reading.
Frahm, Jan-Michael and Koch, Reinhard. Camera Calibration with known Rotations. 2003. in IEEE International Conference on Computer Vision (ICCV).
How a known rotation helps in camera calibration. More detailed description with derived equations is given in Frahm's PhD thesis.